Session H10: Biological Fluid Dynamics: Physiological Phonation and Speech

Phonation aerodynamics estimated from vocal tract acoustics

Measurements in a model of the human upper airway are presented.  The model was comprised of a 2.54cm square duct, with molded rubber vocal folds that divided the duct into trachea and vocal tract sections.  Synchronous measurements of pressure using two Kulite XCS-093 pressure transducers and 5 Larson-Davis ½" microphones, distributed over the length of the duct, and high-frame rate imaging of the glottis were performed for a range of subglottal pressures.  Cross-spectral analysis of acoustic pressure measurements were used to extrapolate acoustic pressure and volume flow throughout the duct.  This information was then used to compute integral quantities such as glottal volume flow, vocal fold drag, and power flows between laryngeal flow and the trachea and vocal tract acoustic fields.   

Computational modelling of simulated unilateral vocal fold paralysis

To improve the voice production of patients with unilateral vocal fold paralysis (UVFP), the thyroplasty procedure (TP) is widely used by inserting an implant surgically into the paralyzed side of the vocal fold (VF) to facilitate VF adduction and vibration. Our cross-disciplinary team has developed a suite of computational tools to simulate the fluid-structure interaction (FSI) of VF vibration, which is coined “PhonoSim”, to improve the TP outcome. This suite includes 3D FEM model of larynx, simplified FSI model with a 1D flow approximation, and high-fidelity 3D FSI model for the VF vibration. We also develop a surgical approach to achieve four different VF configurations: no adduction for open VF, two-side suture to mimic healthy phonation, one-side suture for UVFP, and one-side suture/one-side implant for TP surgery. Ex-vivo MRI scans of four configurations are used to provide the VF geometries. After model reconstructions, a 3D FEM model is utilized to simulate VF adductions, which will provide the adducted geometry and prestress under different conditions for FSI simulations. Simplified FSI simulations are performed for tissue property identification and initial implant design. 3D FSI is utilized to finalize the implant design. Details of these computational components and their combination with the experiment study will be presented.    

Fluid-structure energy exchange during phonation: investigation of dissipated collision energy by direct measurement of internal tissue velocity

The vocal fold (VF) oscillation kinematics during phonation arise due to the balance between the energy provided by the airflow and that dissipated by the VF tissue, with most of the energy dissipated during the collision phase. Elucidating the mechanics of energy dissipation within the tissue is, therefore, paramount to gaining insight into the dynamics of VF oscillations. Assessing the dissipated collision energy requires computing the kinetic energy of the VF immediately preceding contact, which is a function of the  VF tissue velocity. This study aims to quantify the distribution of the tissue velocity inside a synthetic VF model to calculate the associated dissipated collision energy. Measurements were performed with a multi-layer, self-oscillating silicone VF model in a hemilaryngeal flow facility. Fluorescent particles were mixed and cured within the different silicone layers during fabrication, and then illuminated by a laser sheet during self-oscillation. Digital image correlation was performed with an sCMOS camera to compute the interior VF tissue velocity and calculate the dissipated energy. Results revealed that the velocity magnitude decays quickly as a function of distance from the medial surface, and the VF kinetic energy is mainly limited to the cover layer.   

Synchronous high-speed fluid structure interaction in a glottal jet model

Surface deformation of a synthetic silicone hemilarynx in the Penn State Upper Airway Model is acquired synchronously with flow measurement of the pulsatile jet at pressures required to produce phonation in the fold. Stereo particle image velocimetry (PIV) is utilized for flow measurements using two Phantom v1212 high-speed cameras at 1 kHz repetition rate for image pairs illuminated by a simultaneous laser sheet. The surface deformation is captured using digital image correlation (DIC), using three Phantom v1212 high-speed cameras, at a repetition rate of 5 kHz of single images, illuminated by a triggered LED system simultaneously. Three DIC cameras are utilized to capture all sides of the vocal fold. The DIC and PIV image/illumination timing is set so that the DIC LED illumination will not be captured by the PIV cameras, and vice versa via an incoming 25 kHz clock signal to delay the DIC trigger pulse. The proper orthogonal decomposition (POD) is performed on the surface deformation to organize modal motions by energy content and correlate with the fluid motion in the jet.   

Jet dynamics in a scaled up vocal fold model with full and incomplete closure

This study focuses on phonation in the physiological condition where the vocal folds do not fully close. While this occurs naturally in children and adult females, there are pathological conditions which can be problematic. Experiments were conducted using a 10x scaled-up model in a free surface water tunnel. 2-D vocal fold models with semi-circular medial surfaces were driven inside a square duct with constant opening and closing speeds. Cases where the vocal folds closed fully and to only 15% of the maximum gap were examined. Time resolved DPIV and pressure measurements along the duct centerline were made at Re = 7200 over equivalent life frequencies from 52.5 Hz to 97.5 Hz. Phase-averaged analysis of key contributors to sound production was conducted along with a dimensional scaling analysis to examine frequency dependencies between different terms in the streamwise integral momentum equation. When the folds do not fully close, there is always a non-zero mean flow downstream of the glottis which significantly alters the dynamics. Extrapolation to higher frequencies suggests changes in force distributions on vocal folds as unsteady inertia becomes of the same order as the transglottal pressure force. Implications on energetics and sound quality are explored.   

Control volume analysis of laryngeal aerodynamics based on high-fidelity aeroelastic-aeroacoustic simulation with turbulence model

In this talk, a control volume analysis of laryngeal airflow is performed from high-fidelity aeroelastic-aeroacoustic simulations. The fully-coupled simulations use the immersed finite element method with slightly compressible fluid formulation to properly resolve acoustics and the Spalart-Allmaras turbulence model to account for subgrid turbulence. Vocal folds mimic the swept-ellipse multilayer rubber model used in coordinated experiments. Simulations were run for a range of subglottal pressures. Terms of the Integral equations for momentum and mechanical energy are computed in the laryngeal control volume to estimate the acoustic source strength and energy utilization in phonation.   

Effects of Vibration on Small Blood Vessel Perfusion within the Vocal Folds

Vibration of small blood vessels in the human body has been shown to vary perfusion flow rate, which in turn influences processes such as oxygen, nutrient, and waste transport. In particular, the vasculature within the human vocal folds is subjected to sustained, high-frequency, large-amplitude vibrations. Consequently, an improved understanding of how vibration characteristics affect perfusion rate within the vocal fold vasculature is sought. To this end, a study of a three-dimensional computational model of flow through a vibrating tube was undertaken. The flow, physical, and vibratory characteristics were based on previous experiments and defined so as to simulate the nature of flow through a vibrating vocal fold. In this presentation, the setup for the computational model is described. Time step, grid size, and convergence criteria studies used to verify the model are summarized, as are the comparisons of simulation output to prior experimental results for model validation. Subsequent investigations to quantify the relationships between perfusion rate and input variables such as vibration amplitude and frequency, vessel diameter and length, and flow properties are presented. These results are used to draw conclusions about the influence of vocal fold vibration on perfusion flow rate, and implications for voice production and anatomical health are discussed.   

Flow-structure interaction simulation of a sei whale larynx

Airflow within the respiratory system is considered the sound source of the vocalization of baleen whales. While the U-fold in the baleen whale larynx has been suggested as a mammal vocal fold homolog, the sound production mechanism is still not clear. In the current study, thanks to a rare opportunity of obtaining a sei whale larynx, we use a numerical approach to explore the sound production mechanism in baleen whales. In our simulations, the realistic geometries of the sei whale laryngeal components, including the U-fold, the related cartilages, the respiratory tract, and the laryngeal sac, will be reconstructed from CT scans. The flow-structure interactions between the airflow and U-fold tissues will be simulated by using a coupled sharp-interface-immersed-boundary method based incompressible flow solver and finite element based solid mechanics solver. High-fidelity results of 3D vibrations of the U-fold tissues and airflow dynamics in both ingressive and egressive flow conditions will be obtained. Comparison with experimental measurements will be provided. This study will improve the fundamental understanding of the basic sound production mechanism and the possible sound control approaches in baleen whales.   

Effect of Subglottic Stenosis Severity on Voice Production in Realistic Laryngeal and Airway Geometries Using Fluid-Structure-Acoustics Interaction Simulation

In this study, the effect of subglottic stenosis (SGS) on glottal flow dynamics, vocal fold vibration, and acoustics during voice production is examined in realistic laryngeal and airway geometries reconstructed from MRI scans and by employing an in-house three-dimensional fluid-structure-acoustics interaction numerical solver. The preliminary results show that the SGS severities lower than 50% have a negligible effect on voice production. By increasing the SGS severity, the effect became more important for 75% and 90% SGS severities, respectively. The area ratio of the glottis to SGS airway was found to be important. For higher SGS severities, the area ratio is higher, leads to a higher flow resistance and consequently, a higher effect on voice production.   

A computational framework for patient-specific surgical planning of type I thyroplasty

In the present study, a patient-specific computational framework has been proposed to determine the shape of the implant that results in the optimal acoustic and aerodynamic outcomes of the type I medialization surgery for unilateral vocal fold paralysis (UVFP) treatment. The framework combines a three-dimensional high-fidelity continuum model of fluid-structure-acoustics interactions (FSAI) of vocal fold vibrations, a laryngeal muscle mechanics model of vocal fold pre-phonatory posturing, and a genetic algorithm for optimization. The individual control of each muscle provides the opportunity to generate healthy and unilateral-paralyzed laryngeal postures through symmetric and asymmetric activations of the laryngeal muscles, respectively. The implant is virtually inserted into the paretic vocal fold in the cases representing UVFP to mechanically correct the laryngeal postures. The optimization is coupled with FSAI simulations to determine the implant shape that produces the best aerodynamic and acoustic features. The preliminary results show that the optimized implant shapes improved the aerodynamic and acoustic features of the voice compared to the UVFP case.